Species delimitation and biogeography of Adelgidae (Hemiptera): phylogeographic origin of newly discovered Adelges in Taiwan and their North American sister

Abstract

Phylogeographic analyses provide valuable insights for species delimitation and taxonomic decision-making. The family Adelgidae (infraorder: Aphidomorpha) exhibits relatively low species diversity, comprising approximately 63 species primarily distributed across temperate regions. However, the taxonomy of this family remains debatable because of its complex life cycle and high morphological plasticity. The DELINEATE program offers a statistical framework that integrates multiple species concepts and prior taxonomic knowledge to improve species delimitation. In this study, we validated the taxonomic status of 3 newly discovered Adelges species—Adelges breviacus sp. nov., Adelges baborinisanensis sp. nov., and Adelges xueshanensis sp. nov.—and elucidated their phylogeographic characteristics. Our findings indicated that the most recent common ancestor of these 3 species diverged from their North American sister—Adelges lariciatus—approximately 4.6 million years ago and persisted in the coastal mountain regions of southeast China. Subsequently, these 3 adelgids migrated to Taiwan with Picea morrisonicola 1–3 million years ago, when environments became favorable for both Adelges and their host Picea. Ancestral area reconstruction revealed that the origins of the crown groups of Adelges and Picea were associated with the biogeographic connection between East Asia and North America, corresponding to dispersal of Picea from North America to East Asia via the Bering Land Bridge. Although transoceanic dispersal might have contributed to the origin of the crown group of Pineus, current data sets are insufficient to test this biogeographic hypothesis. Overall, long-distance dispersal appears to have facilitated the disjunct distribution and current biogeographic patterns of Adelgidae.

Introduction

Studying speciation through phylogeography improves the understanding of species diversification as well as of the associations between microevolutionary processes and macroevolutionary patterns (Li et al. 2018, Harvey et al. 2019, Fenker et al. 2021). Phylogeographic studies trace the spatial genetic divergence histories of populations or closely related species, offering informative evolutionary frameworks for species delimitation and taxonomic decision-making (Beheregaray and Caccone 2007, Carstens et al. 2013, Garrick et al. 2015, Huang 2020, Smith and Carstens 2020, Chou et al. 2021). Cataloging species diversity is a prerequisite for evolutionary biology research, conservation strategy development, and pest management (de Queiroz 1998, 2005, Mace 2004, Bickford et al. 2007). However, species delimitation is challenging because of inconsistencies across data types and species concepts, such as molecular and phenotypic evidence (Pyron et al. 2016, Schlick-Steiner et al. 2010, Huang and Knowles 2016). For example, cryptic taxa, which are genetically different but morphologically identical, are sometimes identified solely through mitochondrial loci; this approach complicates their taxonomic classification (Dasmahapatra et al. 2010, Havill et al. 2016, Thielsch et al. 2017, Chan et al. 2020, Makhov et al. 2021). Delimitation is also challenging in species with high morphological plasticity (Xiao et al. 2010, Tsai et al. 2014, Lee et al. 2016). Species are regarded as genetic and evolutionary entities and modern systematics relies on molecular approaches to elucidate the historical processes shaping genetic variations.

Phylogeographic and species delimitation studies, which are typically conducted using the same data sets and natural systems, address 2 complementary concerns regarding the speciation continuum: the process of species diversification and the number of resultant species (Huang 2020). Therefore, insights from phylogeographic studies can inform effective species delimitation (Beheregaray and Caccone 2007, Huang 2020). Island biodiversity is predominantly derived from mainland and adjacent islands (Carlquist 1974, Paulay 1994); thus, insular organisms constitute a valuable system for exploring phylogeography-based speciation processes pertaining to isolation and colonization histories (Losos and Ricklefs 2009), such as within-island speciation (Esselstyn et al. 2009, 2013, Tseng et al. 2015, Demos et al. 2016, Huang et al. 2018, Phillips et al. 2020) and back-colonization to mainland (Allan et al. 2004, Filardi and Moyle 2005, Tseng et al. 2015, Takahashi et al. 2021). However, genetic patterns shaped by diverse phylogeographic histories can be complex and incongruent with preexisting taxonomic systems (Yu et al. 2014, Tominaga et al. 2015, Wang et al. 2017, Tseng et al. 2018, Yang et al. 2018, 2021). Understanding the phylogeographic histories underlying speciation can facilitate accurate species delimitation (Beheregaray and Caccone 2007, Huang, 2020).

The Adelgidae family (Hemiptera) comprises approximately 63 species, which are primarily distributed across boreal and temperate regions of the Northern Hemisphere (Favret et al. 2015, Havill et al. 2021). Although Adelgidae has relatively low species diversity, its taxonomy remains debatable because of its complex sexual–asexual life cycle, host alternation, conserved interspecific morphology, and high intraspecific morphological plasticity across life stages (Foottit and Mackauer 1980, Havill and Foottit 2007, Havill et al. 2007, 2021, Foottit et al. 2009, Žurovcová et al. 2010). Traditionally, species boundaries in Adelgidae have been delineated using morphological and ecological data (Steffan 1962, Eichhorn 1967, Havill et al. 2006, Havill and Foottit 2007). With the advent of DNA barcoding (Hebert et al. 2003a, 2003b), DNA barcode-based methods have increasingly been used to determine taxonomic status (Foottit et al. 2009, Žurovcová et al. 2010). Thus, multiple species concepts, such as morphological (or typological), ecological, biological, and phylogenetic (or genealogical) concepts, have been applied to define species limits in Adelgidae. Regarding ecological adaptation, adelgids exhibit a unique life cycle involving host alternation. Adelgids use Picea as their primary host, which supports gall formation and sexual reproduction. They use 5 other Pinaceae genera (Larix, Abies, Tsuga, Pseudotsuga, and Pinus) as secondary hosts, which support parthenogenesis. The life cycle involving an alternation between primary and secondary hosts is called a holocycle, whereas that involving no host alternation is called an anholocycle (Havill and Foottit 2007). Some taxonomists have classified adelgids with different life cycles as separate species, emphasizing their ecological and biological significance (Foottit 1997). For example, holocyclic Adelges viridis and anholocyclic Adelges abietis exhibit minimal morphological differences and negligible genetic divergence; however, because it lacks the sexual stage essential for completing a holocycle, A. abietis is reproductively isolated from A. viridis (Cholodkovsky 1915, Havill et al. 2007). Notably, a single adelgid species can simultaneously exhibit different life cycles across regions, depending on the availability of suitable local hosts for host alternation (Sano et al. 2008, Sano and Ozako 2012, Havill et al. 2014, 2016). These complex ecological adaptations obscure species boundaries, fueling debates over ecologically defined species (Havill et al. 2007, 2021, Foottit et al. 2009, Žurovcová et al. 2010).

Two Adelges species have been documented in subtropical Taiwan, one of which remains taxonomically unnamed (Chen et al. 2014). Taiwan’s landscape is characterized by steep mountains, with >250 peaks exceeding a height of 3,000 m. This environment provides a suitable habitat for highland or temperate organisms, including adelgids and their host plant Picea morrisonicola. Phylogeographic studies have unveiled speciation patterns in Taiwan, driven by colonization and subsequent isolation due to periodic glacial events, as observed in beetles (Huang and Lin 2010, Sota et al. 2011, Weng et al. 2016) and birds (Chu et al. 2013, Dong et al. 2020). Adelges tsugae is a taxonomically identified adelgid species in Taiwan. This species likely originated from mainland China and then spread to Japan; its Japanese population eventually colonized North America (Havill et al. 2016). Notably, A. tsugae has several geographically distinct mitochondrial lineages, which were previously regarded as cryptic species (Havill et al. 2006); however, current taxonomy treats them as a single species. Chen et al. (2014) reported another genetically distinct, unnamed Adelges species characterized by elongated galls, unlike the spherical galls of A. tsugae. Our field observations and analyses revealed that this unnamed species produces galls of various shapes, which suggests hidden species diversity or population differentiation. In some countries, adelgids are recognized as forest pests and invasive species, responsible for high rates of mortality in conifer trees (Smith and Coppel 1957, Hollingsworth and Hain 1991, McClure 1992, Havill et al. 2006, Lazzari and Cardoso 2011, Limbu et al. 2018). Accurate species identification is crucial to effective pest management (Foottit et al. 2009, Havill et al. 2021). Adelgidae’s biological complexity and ongoing taxonomic debates highlight the need for an integrative species delimitation study to determine the taxonomic status of Adelgidae and the phylogeographic history of its putative lineages.

The multispecies coalescent (MSC) model has been widely used for species delimitation. However, MSC models cannot incorporate prior taxonomic knowledge and often misclassify population-level genetic structures as distinct species (Sukumaran and Knowles 2017, Huang 2018, Leaché et al. 2019, Sukumaran et al. 2021). By contrast, the DELINEATE program offers a species delimitation framework that incorporates diverse perspectives on species boundaries (Sukumaran et al. 2021). DELINEATE uses prior taxonomic knowledge to infer unknown taxa, enhancing the accuracy of species delimitation by modeling the extended speciation process and estimating the speciation completion rate to differentiate between population- and species-level genetic structures. This program can clarify how microevolutionary processes shape macroevolutionary patterns (Sukumaran et al. 2021). For DELINEATE analysis in the present study, we referenced 3 studies on the species delimitation of Adelgidae (Chen et al. 2014, Havill et al. 2016, 2021). Chen et al. (2014) reported on unknown Taiwanese adelgids which were genetically distinct from known A. tsugae. Havill et al. (2016) indicated that A. tsugae comprises several divergent mitochondrial lineages with taxonomic uncertainty. Havill et al. (2021) further validated the independent species status of A. piceae and A. nordmannianae, which cannot be differentiated on the basis of mitochondrial loci alone. By leveraging prior knowledge regarding other adelgids, DELINEATE facilitates a statistical evaluation of species delimitation and speciation in adelgids.

In this study, we used data pertaining to 1 mitochondrial locus and 1 nuclear locus to identify the biogeographic patterns of Adelgidae and describe 3 new Adelges species: Adelges breviacus sp. nov., Adelges baborinisanensis sp. nov., and Adelges xueshanensis sp. nov. We hypothesized that periodic glacial events, which led to land bridge formation between Taiwan and mainland Asia, contributed to the colonization of these adelgids and their host plants into the island of Taiwan. To better understand the speciation process, reconstructed the ancestral areas of Adelgidae and inferred the phylogeographic history of Taiwanese adelgids to identify the global and local dispersal and speciation patterns of Adelgidae. On the basis of this framework, we clarified the biogeography of Adelgidae and elucidated mechanisms underlying the current species diversity within this family.

Materials and Methods

Sampling

To sample the Taiwanese adelgids, galls and fundatrices were collected from Picea morrisonicola from 10 localities (Fig. 1). Mature galls were maintained in separate plastic jars until the emergence of winged gallicolae; immature galls were dissected to collect nymphs. For DNA analysis, all gallicolae from the same gall were collectively preserved, and fundatrices were separately preserved in 95% ethanol at −20 °C. Given the parthenogenetic nature of these organisms, only one individual was selected from each gall for DNA extraction. For morphological examination, some gallicolae and fundatrices were processed into slide specimens by using Hoyer’s medium. For genetic analysis, the mitochondrial cytochrome oxidase I (COI) and nuclear elongation factor-1α (EF1α) sequences of other adelgids (Havill et al. 2007, Foottit et al. 2009, Ravn et al. 2013, Žurovcová et al. 2010, Havelka et al. 2020) were downloaded from GenBank. This data set is currently the most comprehensive available for adelgids. However, the COI haplotypes in this data set are numerous and highly redundant. Considering the informative phylogenetic signals and our limited computational resources, we analyzed adelgids with data available for both loci and all EF1α sequences and retained only one sequence for each COI haplotype unless an individual also had EF1α data to reduce the influence of missing data. We excluded adelgids with short sequence lengths and without taxonomic identification. For the newly sequenced taxa, the GenBank accession numbers for COI and EF1α are LC761574–LC761599 and LC761801–LC761826, respectively. All sequences analyzed in this study are presented in Supplementary Table S1.

DNA Extraction, Sequencing, and Alignment

Genomic DNA was extracted from the head and thorax tissues of gallicolae by using the QuickExtract DNA extraction kit (50 μl solution; Epicentre Biotechnologies, Madison, WI, USA). The entire body of each fundatrix was submerged in 50 μl of extraction buffer and incubated at 65°C for 10 min and then at 98 °C for 2 min. After incubation, the fundatrix body was removed, and DNA was extracted. The resultant DNA solution was stored at −20 °C for subsequent polymerase chain reaction (PCR). COI and EF1α sequences were amplified using the primer pairs LCO1490 (5′GGTCAACAAATCATAAAGATATTGG3′)/HCO2198 (5′TAAACTTCAGGGTGACCAAAAAATCA3′) (Folmer et al. 1994) and AdelEF1F1 (5′GTACATCCCAAGCCGATTGT3′)/AdelEF1R2 (5′CTCCAGCTACAAAACCACGA3′) (Havill et al. 2007), respectively. PCR was performed in a reaction volume of 25 μl under the following conditions: initial denaturation at 94 °C for 2 min; followed by 35 cycles of denaturation at 94 °C for 20 s, annealing at 45 °C for COI and 50 °C for EF1α for 40 s, and extension at 72 °C for 45 s; and final extension at 72 °C for 10 min. PCR products were purified using shrimp alkaline phosphatase/exonuclease I (USB Products; Affymetrix, Cleveland, OH, USA). DNA sequences were obtained through thermocycle sequencing by using the BigDye terminator 3.1 sequencing kit (Applied Biosystems, Foster City, CA, USA). The sequences were analyzed using an ABI 3730XL DNA analyzer (Applied Biosystems, Foster City, CA, USA). Sequence quality was evaluated on the basis of electrograms obtained using BioEdit v.7.0 (Hall 1999). Because both COI and EF1α are protein-coding genes, sequences from these loci were aligned using MACSE v.2.05 with default settings (Ranwez et al. 2018) and then edited in BioEdit. Introns in the aligned EF1α sequences were removed before subsequent analyses (Havill et al. 2007).

Phylogenetic Tree Reconstruction

Because of heterozygosity, the EF1α sequences were phased before analysis by using the Phase algorithm in DnaSP v.6.12 (Rozas et al. 2017). The best-fit substitution model for each locus was identified using jModelTest v.2.1.10 (Darriba et al. 2012), leveraging the Akaike information criterion (AIC). COI and EF1α gene trees were constructed using BEAST v.2.7.6 (Bouckaert et al. 2019). A strict clock model was adopted in BEAUti. The clock rate for COI was set at 0.0177 per million years (Papadopoulou et al. 2010), and that for EF1α was set at 1.0 for estimation. The partitioned substitution models for each locus were selected on the basis of the jModelTest results for both COI and EF1α. The remaining parameters for the reconstruction of the COI and EF1α trees were as follows. In the prior panel, the tree model was aligned to the birth–death model, which accounts for extinction and unsampled lineages (Yang and Rannala 1997, Ho and Duchêne 2014, Nattier et al. 2021). By referencing the phylogeny inferred by Havill et al. (2007), we constrained the genera Adelges and Pineus to be reciprocally monophyletic. The times of divergence between Adelges and Pineus and between A. piceae and A. nordmannianae were set to 87.18 (Havill et al. 2007) and 0.1195 (Havill et al. 2021) million years ago (Mya), respectively. The clade age of A. tsugae was set to 0.3231 Mya (Havill et al. 2016). These time points served as references for calibrating the BEAST2 model. Default settings were used for the remaining parameters. The Markov chain Monte Carlo (MCMC) simulation was run for 10⁸ generations at a sampling frequency of 5,000 generations. The output log file was analyzed to determine effective sample sizes (ESS), ensuring values of > 200, by using Tracer v.1.7.1 (Rambaut et al. 2018). The output tree file was processed in TreeAnnotator v.2.7.6 to generate a maximum credibility tree; the common ancestor height was calculated after the first 10% of generations were discarded as burn-in.

Considering that the discordance between the COI and EF1α gene trees could influence our interpretation of phylogenetic data (Fig. 2), we used the Species Tree And Classification Estimation, Yarely (STACEY) MSC model (Jones 2017) for inferring species trees. STACEY accounts for incomplete lineage sorting. It can infer a species tree without requiring species assignment (Jones et al. 2015, Jones 2017). With reference to the STACEY manual, we unchecked the option to estimate collapse weight and set its initial value to 0 for the species tree analysis. Partitioned site models for COI and EF1α were specified on the basis of the results derived from jModelTest. A strict clock was applied with a clock rate of 0.0177 for COI. The ploidy levels for COI and the phased haplotypic EF1α were set to 0.5 and 1.0, respectively. A lognormal prior was applied to the most recent common ancestor (MRCA) priors for both Adelges and Pineus as monophyletic constraints. Other parameters were left as default. An MCMC run of 10⁸ generations was conducted at a sampling frequency of 5000. Tracer v.1.7.1 was used to confirm an ESS value of > 200 for each estimated parameter. The run was resumed to reach an ESS value of > 200, resulting in a total MCMC run of 2 × 10⁸ generations.

Finally, a species tree was generated by summarizing data in TreeAnnotator v.2.7.6 after a 10% burn-in was applied. This STACEY species tree was used to reconstruct a dated phylogeny of Adelgidae. First, a concatenation analysis was performed on the basis of the COI and phased EF1α sequences, with the BEAUti configuration used for the reconstruction of gene trees, to infer the clade ages of the crown groups of Adelges and Pineus in BEAST2. The obtained 95% highest posterior density (HPD) values for the ages of the crown groups of Adelges (10 to 17 Mya) and Pineus (5 to 8 Mya), age of the MRCA (87.18 Mya), age of A. tsugae (0.3231 Mya), and clade age of A. nordmannianae/A. piceae/A. prelli (0.1195 Mya) were used to calibrate the STACEY species tree by using the R package rBt (Sánchez-Ramírez 2018) and the R function chronos() in the package ape (Paradis et al. 2004).

A topology-fixed dating analysis was performed using the calibrated STACEY species tree in BEAST2. In BEAUti, the Site Model and Clock Model panels for COI and EF1α were configured identically to those used in the reconstruction of gene trees. In the priors panel, 3 calibration nodes referenced from other studies were applied—the mean clade ages of A. tsugae (0.3231 Mya), A. nordmannianae/A. piceae/A. prelli (0.1195 Mya), and the crown group of Adelgidae root (87.18 Mya)—by using normal distribution priors with S values of 0.01, 0.001, and 1.0, respectively. In addition, the MRCA priors for the crown groups of Adelges and Pineus were specified with normal distribution priors (M = 13.7 [S = 2.0] and 6.3 [S = 1.0], respectively) to incorporate the previously estimated 95% HPD values. To fix the topology, the calibrated STACEY species tree was used as the starting tree, and the operator weight for the Wide exchange, Narrow exchange, Wilson-Balding, and Subtree-slide operators was set to 0 to maintain the tree throughout the BEAST2 run. Other parameters were set as default. Finally, an MCMC run of 10⁸ generations was conducted at a sampling frequency of 5,000 generations. Tracer v.1.7.1 was used to confirm an ESS value of >200 for each estimated parameter. The final dated tree was generated in TreeAnnotator v.2.7.6 by identifying the best-supported maximum credibility tree, with the common ancestor height calculated after applying a 10% burn-in.

Because Larix-associated adelgids, including the 3 new Taiwanese adelgids, constituted the focal taxon in our study for further phylogeographic inference, we reconstructed a species tree only for these adelgids (Fig. 3A) by using StarBEAST2 (Ogilvie et al. 2017). COI and the phased EF1α sequences of the Larix-associated clade were exclusively analyzed in jModelTest v.2.1.10 to identify the best-fit substitution models. A strict clock and the birth–death model were applied. The clock rate for COI was set to 0.0177, whereas that for EF1α was set to 1.0. The ploidy levels for both loci were set identically to those in the STACEY analysis. The remaining parameters were set as default. An MCMC run was conducted for 10⁸ generations at a sampling frequency of 5,000 generations. Tracer v.1.7.1 and TreeAnnotator v.2.7.6 were respectively used to assess the ESSs of the estimated parameters (>200) and generate the final maximum credibility tree after a 10% burn-in.

Species Delimitation

The DELINEATE program requires a population-level tree, with each tip representing a population (lineage). For population assignment, both the COI and EF1α gene trees inferred through the BEAST2 analysis were analyzed separately using a single-locus species delimitation program, multirate Poisson Tree Processes (mPTP; available at: https://mptp.h-its.org/#/tree; Kapli et al. 2016). A population-level tree was generated by retaining one individual of each lineage on the dated tree in accordance with genetic clusters identified in both the COI and EF1α trees. Other individuals on the dated tree were removed using the R function drop.tip() in the package ape. On the basis of the current taxonomy of Adelgidae, one individual from each species of the following groups was retained despite the mPTP model not delimiting them as separate lineages: Adelges laricis/Adelges tardus, A. abietis/A. viridis, A. nordmannianae/A. piceae/A. prelli, and Pineus orientalis/Pineus pini. Following the DELINEATE guidelines (https://jeetsukumaran.github.io/delineate/index.html), we assigned and constrained the aforementioned lineages as distinct species on the basis of the current species taxonomy. The newly discovered Taiwanese lineages and species exhibiting nonmonophyly on the population-level tree (Adelges japonicus, Pineus coloradensis, and Pineus cembrae) were left unconstrained to determine their species status.

Ancestral Origin Estimation

The ancestral area of Adelgidae was deduced using the R package BioGeoBEARS v.1.1.2 (Matzke 2013a, 2013b). The deduction was performed using the same population-level tree as that used in the DELINEATE analysis, which retained one individual per lineage for population-level representation. On the basis of the native distributions documented by Havill et al. (2007) and Havill and Foottit (2007), we identified 3 geographical regions by their proximity (Fig. 4A): Western Eurasia, East Asia, and North America. The following 3 models were evaluated in BioGeoBEARS by using the AIC and Akaike weights: the dispersal extinction cladogenesis (DEC) model (Ree et al. 2005), likelihood version of dispersal–vicariance analysis (DIVALIKE) model (Ronquist 1997), and likelihood version of the BayArea (BAYAREALIKE) model (Landis et al. 2013). Dispersal multipliers were set on the basis of connectivity among the geographic regions. Furthermore, dispersal multipliers set to 1.0 across all geographic regions were used as null models. The following 6 models were evaluated: DEC_null, DEC, DIVALIKE_null, DIVALIKE, BAYAREALIKE_null, and BAYAREALIKE.

The following 5 periods were defined on the basis of the formation of land bridges (Sanmartín et al. 2001): 0 to 1.5, 1.5 to 3.5, 3.5 to 20, 20 to 30, 30 to 35, and 35 to 90 Mya (Fig. 4). The Bering Land Bridge formed 0 to 1.5 to 3.5 to 90 Mya. Land bridges were available for both Western Eurasia–America and Asia–America. However, Western Eurasia and East Asia were isolated by the Turgai Strait 30 to 90 Mya. Western Eurasia and Asia were disconnected but Asia and America were connected 30 to 35 Mya. On the basis of the connectivity of continents for dispersal, multipliers for well-connected and well-separated regions were set to 1.0 and 0.1, respectively. Although the Bering Land Bridge was available 20 to 30 and 30 to 35 Mya, the multipliers for these 2 periods were set to 0.5 because most dispersal events are believed to have ceased (Sanmartín et al. 2001). Given the long distance between Western Eurasia and North America, corresponding multipliers were set to 0.5 only when the Bering Land Bridge connected Eurasia and North America. A relevant study reported few dispersal events when East Asia and North America were disconnected (Havill et al. 2016). Thus, the multipliers for the connection between East Asia and North America and that between East Asia and Western Eurasia were set to 0.1 to account for long-distance dispersal events during periods when these regions were disconnected.

Results

Molecular Phylogeny of Adelgidae

Although the COI and EF1α trees exhibited similar species clusters, they differed in topology across species-level phylogenetic relationships, indicating gene tree discordance (Fig. 2). Most species-level clusters were well-supported (posterior probability [PP] = 1.0) in both trees; however, some taxonomically valid species were not monophyletic. For example, A. abietis and A. viridis as well as A. nordmannianae and A. piceae were nested within their respective species pairs in both the COI and the EF1α tree. Furthermore, A. japonicus formed a single clade on the COI tree but was polyphyletic and nested with A. laricis and Adelges viridanus on the EF1α tree. In the STACEY tree (Fig. 3A), species associated with the same genus of putative secondary host plants were clustered within monophyletic lineages. The Tsuga-associated clade (A. tsugae) was sister to other Adelges species (PP = 1.0). The Pseudotsuga-associated clade (Adelges cooleyi) and the Abies-associated clade formed a sister group (PP = 0.99). The Larix-associated clade was sister to the combined clade of the Pseudotsuga-associated and Abies-associated lineages (PP = 0.93). The 3 Taiwanese adelgids formed a monophyletic group with low support (PP = 0.59) and appeared to be sister to the North American species A. lariciatus (PP = 1.0). The species tree inferred by StarBEAST2 supported the same topology, with higher PPs (0.9 for the monophyly of the 3 Taiwanese adelgids and 0.95 for the sister relationship with A. lariciatus). These findings are presented in Fig. 3B. The StarBEAST2 tree generally aligned with the topology of the Larix-associated clade in the STACEY tree, but the phylogenetic position of A. japonicus was noteworthy (Fig. 3A and B). Although A. japonicus, A. laricis, and A. tardus formed a well-supported clade (PP = 0.88), the STACEY tree exhibited 2 A. japonicus lineages—one sister to A. viridanus and the other sister to the A. laricis/A. tardus clade—with low support values. In comparison, the StarBEAST2 tree indicated A. japonicus to be sister to A. viridanus, with medium support (PP = 0.79). Molecular dating revealed that the divergence between the 3 new Taiwanese adelgids and A. lariciatus occurred approximately 4.6 (95% HPD: 3.74 to 5.48) Mya, whereas the clade age of the 3 Taiwanese taxa was 4.03 (95% HPD: 3.22 to 4.86) Mya. The crown groups pf Adelges and Pineus originated approximately 11.7 (95% HPD: 9.36 to 14.12) and 5.5 (95% HPD: 4.49 to 6.64) Mya, respectively.

Species Delimitation Revealed 3 New Adelges Species in Taiwan

The mPTP analyses identified 20 and 24 lineages on the EF1α and COI trees, respectively, including 3 newly described lineages from Taiwan, which generally aligned with the current species taxonomy (Fig. 2). However, the analyses did not differentiate among the following taxonomically recognized species groups: A. laricis/A. tardus, A. abietis/A. viridis, A. nordmannianae/A. piceae/A. prelli, and P. orientalis/P. pini. The mPTP analyses further unveiled intraspecific divergent lineages for A. pectinatae, A. cooleyi, and P. cembrae. Notably, previously identified lineages of A. tsugae were not detected in the mPTP analysis for COI or EF1α.

For the 3 new Taiwanese lineages, both species delimitation models—mPTP and DELINEATE—corroborated species-level divergence for A. breviacus, A. baborinisanensis, and A. xueshanensis (Figs. 2 and 3C). Among the unconstrained lineages, DELINEATE delimited the 2 lineages of A. japonicus as distinct species. Two of the 3 lineages of P. cembrae were delimited as separate species. Furthermore, one lineage of P. coloradensis was delimited as a distinct species, whereas the other lineage was conspecific with one of the P. cembrae lineages (Fig. 3C).

Historical Biogeography of Adelgidae

Table 1 presents the BioGeoBEARS results for the 6 aforementioned biogeographic models. Among these models, the DEC_null and DEC models obtained the lowest AIC score (log-likelihood = −59.04; AIC = 122.1; AIC weight = 0.29) and were thus identified to be the best-fitting models for the empirical data. These results suggest that dispersal multipliers exerted no significant effect on model testing in our case. The origin of Adelgidae was associated with the connection of East Asia and North America via the Bering Land Bridge. The extant Adelges likely originated in this East Asia–North America region, whereas the extant Pineus likely originated in the Western Eurasia–North America region. The extant Adelges primarily colonized East Asia and Western Eurasia; exceptions to this include A. cooleyi and A. lariciatus, which colonized North America, and A. tsugae, which originated in Asia and later colonized North America. The ancestral lineage of the crown group of Pineus was likely widespread across the Western Eurasia–North America region and subsequently colonized North America and Eurasia (Fig. 4). BioGeoBEARS analyses further indicated that the ancestral region for the 3 new Taiwanese taxa and A. lariciatus was associated with the connection between East Asia and North America.

Discussion

In this study, we examined the phylogeographic and speciation patterns among closely related species up to broader phylogenetic levels. This helped us identify the overall biogeographic pattern for Adelgidae. We further discovered 3 new adelgids native to Taiwan. Integrated analyses of phylogeographic, historical biogeographic, and ecological data suggested that the new adelgids and their North American sister, A. lariciatus, can be used to establish a novel system for studying the comparative phylogeography of Adelges and Picea. Our biogeographic inference revealed a strong association between adelgids and their host plants.

Species Delimitation of and Ecological Insights from the New Taiwanese Adelgids

The DELINEATE analysis combined with additional ecological data supports taxonomic recognition of the 3 Taiwanese adelgids as distinct species. These adelgids exhibit some degree of co-occurrence, including in the Xueshan area. Phylogenetically, they belong to the Larix-associated clade, which indicates Larix as their secondary host (Havill et al. 2007). However, because Larix is not native to Taiwan, the 3 Taiwanese adelgids are likely anholocyclic. Despite their current sympatric or parapatric distribution, the marked genetic divergence among these taxa and the absence of a secondary host for sexual reproduction indicate that they are separately evolving metapopulation lineages (de Queiroz 2005, 2007); this notion aligns with the unified species concept (de Queiroz 2005). Therefore, A. breviacus sp. nov., A. baborinisanensis sp. nov., and A. xueshanensis sp. nov. are biologically and evolutionarily distinct species requiring formal taxonomic description.

The biogeographic pattern based on the BioGeoBEARS result suggests that a lineage diverging from the common ancestor with A. lariciatus evolved into the MRCA of the 3 new Taiwanese adelgids. This scenario indicates 2 phenomena. First, the MRCA of the 3 new Taiwanese adelgids likely underwent long-distance dispersal from a region near Beringia into East Asia. Second, a common ancestor with the holocyclic adelgid A. lariciatus evolved into the anholocyclic adelgids A. breviacus, A. baborinisanensis, and A. xueshanensis. This hypothesis is consistent with assumptions that anholocyclic adelgids are derived from holocyclic ones (Havill and Foottit 2007, Havill et al. 2007) and that parthenogenetic species are derived from sexual species (Simon et al. 2003). Furthermore, the results of BioGeoBEARS analyses and molecular dating suggest that the MRCA of the 3 Taiwanese adelgids originated in East Asia before the orogeny on the island of Taiwan, which created a suitable environment for adelgids. Therefore, the MRCA of the 3 new Taiwanese adelgids likely existed or still exists on the East Asian mainland, suggesting the presence of extinct or unsampled lineages that are more closely related to the Taiwanese species in East Asia. According to Havill and Foottit (2007), at least 3 unsampled Larix-associated adelgids are present in East Asia and 4 in Europe; however, the distribution of several related species remains unclear. Thus, the species most closely related to the Taiwanese adelgids might not have been included in this study. Nonetheless, the unsampled Larix-associated species exhibit life cycles or host preferences different from those of the Taiwanese taxa, ruling out conspecificity due to differing ecological traits. Most Larix-associated species have been reported in Europe and Japan (Havill and Foottit 2007), although Larix is widespread across the temperate regions of Eurasia. Genetically divergent adelgid species can exhibit poorly distinguishable morphology and diverse gall shapes (Fig. 5). Therefore, the widespread distribution of Larix in temperate Eurasia warrants future investigations into the diversity of Larix-associated Adelges species.

Notably, A. baborinisanensis and A. xueshanensis may have a broader distribution than does A. breviacus. However, whether these 3 species exhibit different phenological features or microhabitat preferences remains unclear. Although Chen et al. (2014) investigated the effect of accumulated temperature on gall formation in Adelges species found near Xueshan (Taichung, Taiwan), the target species could not be clearly identified because of the presence of cryptic species and a lack of accurate identification methods at the time. Moreover, only the fundatrices of A. baborinisanensis have been documented thus far. Therefore, future studies should employ extensive genomic sampling to explore ecological adaptation processes, such as interspecific competition, niche partitioning, and phenology. Furthermore, investigating the effects of periodic glaciation on population dynamics and migration patterns may unravel the speciation processes for these sympatric or parapatric species.

From Phylogeographic Processes to Biogeographic Patterns

Our BioGeoBEARS analyses suggest that the Taiwanese adelgids evolved from a common ancestor that diverged from A. lariciatus. Although the dated phylogenetic tree indicates that the origin of the Taiwanese adelgids postdates the emergence of Taiwan and the divergence of P. morrisonicola from its continental relatives, these adelgids did not inhabit Taiwan immediately but rather arrived much later. The island of Taiwan is located in a subduction collision zone between the Philippine Sea Plate and the Eurasian Plate; it was formed in the late Miocene, approximately 6.5 Mya (Huang et al. 1997). On the basis of generation time assumptions (ranging from 25 to 50 years) or fossil calibration, researchers have estimated different divergence times for the split between P. morrisonicola and its continental relatives: 4 to 8 (Bodare et al. 2013), 6.78 (Zou et al. 2013), and 6.84 (Shao et al. 2019) Mya. The origin of P. morrisonicola seems to align with the emergence of Taiwan. Furthermore, the Taiwanese adelgid ancestor might have diverged from A. lariciatus approximately 4.6 (95% HPD: 3.74 to 5.48) Mya, A. xueshanensis might have diverged from the common ancestor of A. breviacus and A. baborinisanensis approximately 4.03 (95% HPD: 3.22 to 4.86) Mya, and A. breviacus might have diverged from A. baborinisanensis approximately 3.54 (95% HPD: 2.71 to 4.39) Mya when the island of Taiwan had emerged. However, the reconstruction of paleovegetation in the late Miocene challenges this scenario. During the early stages of Taiwan’s formation, environmental conditions were unsuitable for Picea plants because of lower elevations and a warmer climate than those of current conditions (Hao et al. 2017). Coastal mountain regions in southeast China likely provided suitable habitats for Picea during the late Miocene epoch (Hao et al. 2017). The drastic orogeny on the island of Taiwan 1 to 3 Mya (Teng 1990, Huang et al. 1997) might have subsequently created conditions conducive to Picea habitation. Bodare et al. (2013) reported that P. morrisonicola likely arrived in Taiwan during the late Pliocene to Pleistocene glaciations (approximately 1 to 3 Mya), a timeframe consistent with the island’s colonization by other conifer genera, such as Taiwania (Chou et al. 2011), Taxus (Gao et al. 2007), and Cunninghamia (Lu et al. 2001). Therefore, the ancient P. morrisonicola populations in the coastal mountains of southeast China might have expanded eastward and then contracted on the island of Taiwan during glaciation periods. Similarly, the ancestor of the new Taiwanese adelgids may have followed an analogous eastward expansion and contraction pattern during this time.

Specifically, the divergence of the Taiwanese adelgid ancestor from A. lariciatus aligns with that of P. morrisonicola from its continental relatives during the middle Pliocene. Bodare et al. (2013) suggested that P. morrisonicola likely diverged from Picea wilsonii approximately 4 (95% HPD: 2.8 to 5.5) Mya. The sister lineages of P. morrisonicola—P. wilsonii and Picea maximowiczii—are currently distributed across the Tibetan Plateau and northern Asia, and Japan, respectively. Insights from ancestral area reconstructions suggest that P. morrisonicola originated near the Tibetan Plateau (Shao et al. 2019). The ancestral area reconstruction performed in our study revealed an Asian origin for the Larix-associated clade and an East Asian–North American origin for the ancestor of A. lariciatus. This evolutionary scenario suggests that the origin of A. lariciatus was associated with the migration of P. wilsonii to North Asia. Subsequently, a lineage diverging from the ancestral lineage of A. lariciatus evolved into the MRCA of the 3 Taiwanese adelgids from late Miocene to middle Pliocene (95% HPD: 3.74 to 5.48 Mya). Simultaneously, the Taiwanese MRCA and A. lariciatus colonized the coastal mountains of southeast China and North America, respectively. This event involved long-distance dispersal from Northeast Asia or Beringia to southeast China. The extant Taiwanese adelgids subsequently colonized Taiwan with P. morrisonicola through land bridges during the late Pliocene to Pleistocene glaciations. Picea is now extinct in southeast China (Li 1995, Kayama et al. 2002, Yang et al. 2009, Li et al. 2012, Hao et al. 2017). Therefore, the colonization of Taiwan likely represents a vicariance or range shift with host plants. Regarding the East Asian–North American origin of A. lariciatus, its ancestor likely persisted in Northeast Asia near Beringia or migrated to regions where extant or currently extinct Picea and Larix species were distributed; this ancestor might have been holocyclic. Host plants play key ecological roles in shaping the diversity of specialist phytophagous aphids (both globally and regionally), exhibiting a strong positive correlation with species diversity; by contrast, climate influences species diversity only on a broad spatial scale (Du et al. 2020). The origin times of the 3 Taiwanese adelgids predate the orogeny and emergence of suitable environments in Taiwan, suggesting that speciation occurred on the Asian mainland from late Miocene to middle Pliocene. Our biogeographic inference and late Miocene paleovegetation reconstructions offer a possible scenario for how Adelges species migrated with Picea, highlighting long-distance dispersal and the resultant disjunct distribution of sister lineages. Although the 95% HPD values for the times of divergence between P. morrisonicola and its sister species and between the Taiwanese MRCA and A. lariciatus extend into the late Miocene, the mean divergence time of approximately 4 Mya indicates middle Pliocene, a period when Picea distribution was unclear. Further investigations into historical allopatry and ancient Picea distribution are required to elucidate the phylogeographic history of Adelges species in Taiwan.

The 3 newly discovered Taiwanese adelgids and A. lariciatus provide valuable insights into the long-distance dispersal and current disjunct distribution of Adelgidae. This distribution pattern is consistent with the biogeography of other aphids and their associations with host plants (Ren et al. 2013, 2017, Meseguer et al. 2015). Ancestral area reconstruction revealed that some Adelges and Pineus species might have undergone long-distance dispersal events between East Asia and North America and even between Western Eurasia and North America during the period when land bridges were absent (Fig. 4). Given the dependence of Adelgidae on Pinaceae and the assumption that the holocyclic life cycle is ancestral, the historical biogeographic pattern of Adelgidae likely reflects the dispersal history of Picea, the primary host of Adelgidae. Picea originated in North America and experienced at least 4 dispersal events to Eurasia through the Bering Land Bridge (Shao et al. 2019). Two of the dispersal events to Asia, occurring approximately 10.27 and 13.97 Mya, might have contributed to the origin of the crown group of Adelges (11.7 Mya; 95% HPD: 9.36 to 14.12 Mya), which predominantly comprises Asian and European adelgid species. The ancestral area of this crown group appears to coincide with the formation of a connection between East Asia and North America (Fig. 4). By contrast, Pinus, the secondary host of Pineus, likely originated in East Eurasia and experienced several dispersal events between Eurasia and North America (Eckert and Hall 2006). However, the ancestral origin of the crown group of Pineus may be associated with the formation of a connection between Western Eurasia and North America during the period when land bridges were absent (Fig. 4). Therefore, the evolution of the crown group of Pineus likely involves several transoceanic long-distance dispersal events. This hypothesis requires further validation with Pineus data sets because the current data may underrepresent East Asian and Eurasian Pineus species, potentially biasing biogeographic inferences. In addition to local extinction of ancient host conifers, which might have led to the disjunct distribution of extant adelgids, passive dispersal through wind likely played a vital role in shaping the current disjunct distribution of adelgids (McClure 1990, Lass et al. 2014). This mechanism aligns with the observed close phylogenetic relationships among adelgid species across continents (Fig. 4). Migratory birds contribute to the long-distance dispersal of adelgids by transporting eggs or first-instar nymphs, which are active crawlers (McClure 1990, Russo et al. 2016, 2019). For example, A. tsugae, which originated in East Asia, is hypothesized to have colonized western North America through bird-mediated dispersal (Havill et al. 2016). Thus, although adelgids are not strong fliers, long-distance dispersal appears to be a prevalent phenomenon in Adelgidae, facilitating their intercontinental disjunct distributions.

Implications for Systematics and Biogeography of Adelgidae

The species delimitation framework used in this study confirmed the species-level divergence of the 3 newly discovered Adelges species in Taiwan and highlighted the importance of population-level phylogenetic analyses for specific species groups. The current species taxonomy of Adelgidae largely corresponds with the genetic clusters identified through the mPTP model; however, several potential cryptic lineages were also detected. Moreover, our mPTP analyses did not support the cryptic species hypothesis for A. tsugae when a family-level phylogenetic tree was used, warranting further population genetic studies. Another key finding is the presence of 2 polyphyletic lineages within A. japonicus, which were delimited as distinct species by DELINEATE but treated as a single species by mPTP on the COI tree. The monophyly observed on the COI tree eliminates the likelihood of misidentification and indicates gene tree discordance. The discordance between the EF1α and COI trees might have resulted from incomplete lineage sorting or gene flow between species (Maddison 1997). The inferred gene trees suggest gene flow between A. japonicus and A. laricis or A. japonicus and A. viridanus. However, given that A. japonicus and A. viridanus are anholocyclic on Picea and Larix, respectively, incomplete lineage sorting or the retention of ancestral polymorphism appears to be the more plausible explanation. A similar pattern was observed in P. cembrae.

Adelgids serves as a valuable system for studying species delimitation, offering a means of investigating various ecological roles and species concepts. Investigations of the phylogenetics of Adelgidae and the discovery of 3 new Adelges species have created a need for a comprehensive survey of Adelgidae systematics. In the present study, our goal was not to provide a taxonomic revision of Adelgidae but to investigate the associations between closely related species-level phylogeography and family-level biogeographic patterns. The phylogeographic association between the Taiwanese adelgids and their North American sister as well as the historical range dynamics of Picea provide a broader biogeographic picture of Adelgidae evolution. However, the current DNA sequence data sets, which are restricted to a few loci, preclude investigations into finer-scale divergence, whether at the level of populations or at higher levels of phylogenetic relationships. Nevertheless, our study represents an initial application of statistical frameworks for developing testable biogeographic hypotheses for Adelgidae. It further clarifies molecular phylogenetic patterns in Adelgidae, offering refined hypotheses for future systematic and biogeographic research.

Conclusion

We investigated phylogeographic relationships between insular and continental species and proposed a hypothesis regarding the speciation and biogeography of Adelgidae. This study unveiled the roles of host association and long-distance dispersal in shaping species diversity. We discovered 3 new Adelges species in Taiwan on the basis of a speciation-based species delimitation model and subsequently explored the phylogeography of the Taiwanese adelgids and A. lariciatus. Our findings suggest that the colonization history of adelgids was influenced by the phylogeography of Picea and the orogeny on the island of Taiwan. The biogeographic and dispersal patterns of Adelgidae identified in this study provide insights into species diversification within the family. Specifically, the phylogeographic scenario for A. lariciatus and the Taiwanese ancestor (the MRCA of A. breviacus, A. baborinisanensis, and A. xueshanensis) coincides with the divergence of P. morrisonicola from its relative inhabiting the area near the Tibetan Plateau during the middle Pliocene or late Miocene epoch. After orogenic developments in Taiwan, the 3 adelgids colonized the island with P. morrisonicola during the period between the late Pliocene and early Pleistocene. At a higher phylogenetic level, the origin of Adelgidae appears to be associated with the connection of East Asia and North America in the late Cretaceous period. The crown groups of Adelges and Pineus are inferred to have East Asian–North American and Western Eurasia–North American origins, respectively. The historical biogeography of conifers and the potential long-distance dispersal across continents might have shaped the current species diversity of Adelgidae. Broader molecular sampling is required to uncover potential hidden species diversity and reconstruct a robust phylogeny for Adelgidae. In addition, studying ancient conifer range dynamics by using species distribution models may improve the understanding of speciation in Adelgidae and its coevolution with Pinaceae.

Taxonomy

Adelges breviacus sp. nov.

Zoobank:

urn:lsid:zoobank.org:act:2E3C320B-BD5E-473B-ABDD-2CAF89B139C7

Etymology.

The specific name is a compound of the Latin brevis and acus, which mean short and needle, respectively. Adelges breviacus sp. nov. is named for its external morphology of galls, whose needles are comparatively short compared to the galls of the other 2 new species.

Supposed common name.

Short-needle adelgid (短棘球蚜)

Holotype.

Adult gallicola (Fig. 6A; glass slide specimen). Original label: “Location: Xueshan 4K (121.280263, 24.38945)/Date: 2019.VI.27/Collector: Ming-Hsun Chou/Code: A5-1.” Glass slide specimen of holotype is deposited in National Chung Hsing University (Taichung, Taiwan).

Paratypes.

Four adult gallicolae (glass slide specimens; gall code: A5-1): same collecting information as for holotype. Three first-star gallicolae (glass slide specimens; specimen code: A5-3): same collecting information as for holotype. One adult gallicola (glass slide specimen; gall code: A10-7): Taiwan, Xueshan 4K (121.280263, 24.38945), 27 August 2019, coll. Wen-Bin Yeh. Glass slide specimens of paratypes are deposited in National Chung Hsing University (Taichung, Taiwan).

Diagnosis.

The galls are distinguishable from those of A. baborinisanensis and A. xueshanensis. The external surface of fresh gall lacks white scale-like structure, and the needles are visibly shorter than those of A. baborinisanensis and A. xueshanensis, especially in the matured or nearly matured galls. For gallicolae, Lack of visible abdominal lateral wax glands is the most distinguishable difference from A. baborinisanensis and A. xueshanensis, which both have obvious lateral wax glands on their abdomen. The cubitus vein of hindwing of gallicola is nearly perpendicularly connected with the radius vein although sometimes the cubitus vein of hindwing is less obvious (Fig. 7A). The barcoding COI sequence amplified by LCO1490/HCO2198 (Folmer et al., 1994) from the same gall as holotype is provided below.

Sequence of DNA Barcode Using the Mitochondrial Cytochrome Oxidase I (COI) Gene (GenBank Accession Number: LC761595)

ACTATATACTTTTTATTTGGAATATGATCAGGAATAATTGGCTCTTCATTAAGAATTATTATTCGACTAGAATTAAGACAAATCAATTCAATTATTAATAATAATCAACTTTACAATGTTATTATTACAATTCATGCTTTTATTATAATTTTTTTTATAACAATACCTATTGTAATTGGGGGATTTGGTAACTGATTAATCCCTATAATAATAGGATCACCTGATATATCATTCCCACGATTAAATAATATTAGATTTTGATTATTACCCCCCTCATTGATAATAATAATATTTAGATTAATAATTAATAATGGAACAGGAACAGGATGAACAATCTATCCCCCACTATCTAATAATATTGCACATAATAATATTTCAGTAGATTTAACCATTTTTTCATTACATATAGCAGGAATCTCATCAATTTTAGGAGCAATTAATTTTATTTGTACAATTTTAAATATAATACCTAATAATATAAAACTAAACCAGATCCCTCTTTTCCCATGATCAATTTTAATTACCGCAATTTTATTAATCATCTCATTACCTGTTCTAGCAGGTGCTATTACAATATTATTAACAGATCGTAATTTAAATACATCTTTTTTTGACCCCTCAGGAGGAGGAGATCCAATTCTATATCAACATTTAT

Description of gallicola.

Antennae with 5 segments. Head and thorax dark, where wax glands almost absent or very small and unobvious. Abdomen with one spiracle pair on each II-VII abdominal segment. Abdominal lateral wax glands absent or hardly visible. Abdominal middle lateral and spinal wax glands on tergites very small and even sometimes absent. IX abdominal segment bearing ovipositor. Costa, subcostal, and radius veins of forewing light green when alive. Second cubitus vein of forewing connected to radius vein while media and first cubitus veins merely very weakly connected to radius vein, and usually media and radius veins more visible disconnected. Cubitus vein of hindwing short and connected to radius vein while sometimes less visible (Fig. 7A).

Description of first-star gallicola.

Body elliptical. Antennae with 3 segments. Head with one pair of wax gland plates anteriorly. Prothorax, mesothorax, and metathorax with one pair of lateral wax gland plates. Each I-VI abdominal segment with one pair of spinal, middle lateral, and lateral wax gland plates. One pair of wax gland plates present on VII abdominal segment dorsally and laterally while VIII segment only with a pair of lateral wax gland plates. Wax glands on IX abdominal segment absent or invisible (Fig. 6E).

Ecology.

According to our collection, galls with gallicolae can be found on Picea morrisonicola during June to early September. We have also recorded that winged gallicolae emerged from the gall in the ends of June and August, and initial galls can also be found simultaneously in the end of June.

Adelges baborinisanensis sp. nov.

Zoobank:

urn:lsid:zoobank.org:act:836D8726-DB9D-4649-BA93-4CFFDAB393F6

Etymology.

Adelges baborinisanensis sp. nov. is named for Dasyueshan, where the holotype was collected. The Atayal people, a Taiwanese indigenous tribe, call Dasyueshan as Babo Rinisan. Babo means a mountain, and Rinisan means weeping tears in Atayal.

Supposed common name.

Weeping mountain adelgid (淚嶺球蚜)

Holotype.

Adult gallicola (Fig. 6B; glass slide specimen; gall code: A11-2). Original label: “Location: Dasyueshan National Forest Recreation Area (121.008285, 24.256272)/Date: 2019.IX.08/Collector: Ming-Hsun Chou/Code: A11-2.” Glass slide specimen of holotype is deposited in National Chung Hsing University (Taichung, Taiwan).

Paratypes.

Five adult gallicolae (glass slide specimens; gall code: A11-1): Taiwan, Dasyueshan National Forest Recreation Area (121.008285, 24.256272), 08 September 2019, coll. Ming-Hsun Chou. Three adult gallicolae (glass slide specimens; gall code: A11-2): Taiwan, Dasyueshan National Forest Recreation Area (121.008285, 24.256272), 08 September 2019, coll. Ming-Hsun Chou. Three adult fundatrices (glass slide specimens; specimen codes: A17-1, A17-2, and A17-3, respectively): Taiwan, Dasyueshan National Forest Recreation Area (121.008285, 24.256272), 21 March 2020, coll. I-Hsuan Chu. Two first-star gallicolae (glass slide specimens; specimen codes: A17-5,6): 25–26 March 2020 hatched from eggs laid by fundatrices A17-5 or A17-6 collected in Taiwan, Dasyueshan National Forest Recreation Area (121.008285, 24.256272) on 21 March 2020 by Ming-Hsun Chou. Three first-star gallicolae (glass slide specimens; specimen codes: A17-7,8): 25–26 March 2020 hatched from eggs laid by fundatrices A17-7 or A17-8 collected in Taiwan, Dasyueshan National Forest Recreation Area (121.008285, 24.256272) on 21 March 2020 by Daniel Lau. Three adult fundatrices (glass slide specimens; specimen codes: A18-1, A18-2, and A18-3, respectively): Taiwan, Dasyueshan National Forest Recreation Area (24.282995, 121.026976), 21-March-2020, coll. I-Hsuan Chu. One adult fundatrix (glass slide specimen; specimen code: A20-1): Taiwan, Guanyun Villa (121.340353 24.187533), 18 February 2020, coll. Wen-Bin Yeh. One adult fundatrix (glass slide specimen; specimen code: A21-3): Taiwan, Xueshan 4K (121.280263, 24.38945), 13 April 2020, coll. Wen-Bin Yeh. Glass slide specimens of paratypes are deposited in National Chung Hsing University (Taichung, Taiwan).

Diagnosis.

The external surface of fresh gall is with white scale-like structure, and the needles are slender, which can be distinguishable from those of A. breviacus. However, the external morphology of gall is almost identical to that of A. xueshanensis. For gallicolae, disconnection between cubitus and radius veins of hindwings are diagnosable from A. breviacus and A. xueshanensis (Fig. 7B). The barcoding COI sequence amplified by LCO1490/HCO2198 (Folmer et al., 1994) from the same gall as holotype is provided below.

Sequence of DNA Barcode Using the Mitochondrial Cytochrome Oxidase I (COI) Gene (GenBank Accession Number: LC761581)

ACTATATATTTTATATTTGGAATGTGATCAGGAATAATTGGATCCTCATTAAGAATTATTATTCGATTAGAATTAAGACAAATTAATTCAATTATTAATAATAATCAAATTTATAATGTTATCATTACAATTCATGCTTTTATTATAATTTTCTTTATAACAATACCAATTGTTATTGGAGGATTTGGTAATTGATTAATTCCTATGATAATAGGATCCCCCGATATATCATTTCCACGATTAAATAATATTAGATTTTGATTATTACCCCCTTCATTAATATTAATAATATTTAGCTTAATTATTAATAATGGAACAGGGACAGGATGAACAATCTATCCACCATTATCTAATAACATTGCACATAATAACATCTCAGTAGATTTAACTATTTTTTCATTACACATAGCAGGAATTTCATCAATTTTAGGAGCAATTAATTTTATTTGCACAATTTTAAATATAATACCTAATAATATAAAAATAAACCAAATCCCACTTTTCCCATGATCAATTTTAATTACAGCAATTTTATTAATTCTATCATTACCTGTATTAGCAGGAGCTATTACAATACTACTTACAGATCGTAATTTAAATACATCTTTTTTTGACCCTTCAGGAGGAGGTGATCCAATTTTATACCAACATTTAT

Description of gallicola.

Antennae with 5 segments, with one small finger-like protrusion on V antennal segment top. Head dark, with wax gland pairs anteriorly and posteriorly. Pronotum dark, with wax gland plates laterally and posteriorly. Mesonotum and metanotum dark, both with a pair of rounded wax gland plates dorsally. Abdomen dark brownish, with one spiracle pair on each II-VI abdominal segment. Each I-VII abdominal segment with spinal, middle lateral, and lateral wax gland plate pairs while spinal wax glands on VII tergite absent. VIII abdominal segment with one lateral wax gland plate pair and a few posterior wax glands. IX abdominal segment bearing ovipositor. Costa, subcostal, and radius veins of forewing light green when alive. Second cubitus vein of forewing connected to radius vein while first cubitus vein disconnected or merely very weakly connected to radius vein. Media and radius veins of forewing visibly disconnected. Cubitus vein of hindwing disconnected or very slightly connected to radius vein (Fig. 7B).

Description of first-star gallicola.

Body elliptical. Antennae with 3 segments. Wax glands on head, thorax, and abdomen likely absent, or very invisible (Fig. 6F).

Description of fundatrix.

Antennae with 3 segments. Head with 3 pairs of wax gland plates between eyes, 2 in front of antennae, and one on each of lateral sides. Eyes represented by triommatidia. Upside of prothorax with 2 pairs of each spinal and middle lateral wax gland plates, and with one pair of lateral wax gland plates. Upsides of meso- and metathorax with 3 pairs of wax gland plates spinally, middle laterally, and laterally. Underside of prothorax with 2 wax gland plates, and undersides of meso- and metathorax with 4 wax gland plates between legs. Abdomen with one spiracle pair on each II-VI segment. Each of abdominal I-VI segments with 3 pairs of wax gland plates spinally and middle laterally, and laterally while spinal wax glands on V and VI tergites sometimes unobvious or absent. VII segment with one pair of middle lateral and lateral wax gland plates, and only one spinal wax gland plate. Abdominal VIII segment bearing ovipositor, without wax glands (Fig. 6D).

Ecology.

Fundatrices are covered by white wooly wax and can be found on Picea morrisonicola from February to April. We have recorded that winged gallicolae emerged from the gall in early September. In Dasyueshan National Forest Recreation Area (121.008285, 24.256272), A. baborinisanensis is sympatric with A. xueshanensis although the latter seems not common there.

Adelges xueshanensis sp. nov.

Zoobanks:

urn:lsid:zoobank.org:act:C09E3D2D-722A-4916-8375-430C64A43C98

Etymology.

Adelges xueshanensis sp. nov. is named for the locality where the holotype was collected. Xueshan means snowy mountain in Mandarin Chinese.

Supposed common name.

Snowy mountain adelgid (雪山球蚜)

Holotype.

Adult gallicola (Fig. 6C; glass slide specimen; gall code: A14-1). Original label: “Location: Xueshan 4K (121.280263, 24.38945)/Date: 2019.X.17/Collector: Wen-Bin Yeh/Code: A14-1.” Glass slide specimen of holotype is deposited in National Chung Hsing University (Taichung, Taiwan).

Paratypes.

Four adult gallicolae (glass slide specimens; gall code: A12-9): Taiwan, Tataka, 22-September-2019, coll. Wen-Bin Yeh. One adult gallicola (glass slide specimen; gall code: A14-1): same collecting information as for holotype. Two adult gallicolae (glass slide specimens; gall code: A16-1): Taiwan, Dasyueshan National Forest Recreation Area (121.008285, 24.256272), 08 December 2019, coll. Ming-Hsun Chou. Glass slide specimens of paratypes are deposited in National Chung Hsing University (Taichung, Taiwan).

Diagnosis.

The external surface of fresh gall is with white scale-like structure, and the needles are slender, which can be distinguishable from those of A. breviacus. However, the external morphology of gall is almost identical to that of A. baborinisanensis. For gallicolae, the cubitus vein of hindwing is visibly connected with the radius vein (Fig. 7C), and there are few wax glands on the anterior area of the pronotum. The barcoding COI sequence amplified by LCO1490/HCO2198 (Folmer et al., 1994) from the same gall as holotype is provided below.

Sequence of DNA Barcode Using the Mitochondrial Cytochrome Oxidase I (COI) Gene (GenBank Accession Number: LC761584)

ACTATATATTTTATATTCGGAATATGATCAGGAATAATTGGATCCTCATTAAGAATTATTATTCGATTAGAATTAAGACAAATTAACTCAATTATTAATAATACACAACTTTATAATGTTATTATTACAATCCATGCTTTTATTATAATTTTTTTTATAACAATACCTATTGTAATTGGTGGATTTGGAAATTGACTAATCCCAATAATAATAGGAACACCTGATATATCATTCCCACGATTAAATAATATTAGATTTTGATTATTACCCCCTTCATTAATAATAATAATATTCAGTTTAATAATTAACAATGGAACAGGTACAGGATGAACAATTTATCCCCCCTTATCTAATAATATTGCACATAATAATATTTCAGTAGACTTAACTATTTTTTCATTACACATAGCAGGAATTTCATCAATTTTAGGAGCAATTAATTTTATTTGTACAATTTTAAATATAATACCAAATAATATAAAATTAAATCAAATCCCACTTTTTCCATGATCAATTTTAATTACTGCTATTTTATTAATTCTTTCATTGCCGGTACTAGCTGGAGCTATTACAATATTACTTACAGATCGTAATTTAAATACATCTTTTTTTGACCCTTCAGGAGGAGGGGATCCAATTTTATATCAACATTTAT

Description of gallicola.

Antennae with 5 segments. Head dark, with wax gland pairs anteriorly and posteriorly. Pronotum dark, with wax gland plates laterally and posteriorly, and with a few wax glands near anterior area. Mesonotum and metanotum dark, both with a pair of rounded wax gland plates dorsally. Abdomen dark brownish, with one spiracle pair on each of II-VI abdominal segments. Each of I-VII abdominal segments with spinal, middle lateral, and lateral wax gland plate pairs while spinal wax glands on VII tergite sometimes absent. VIII abdominal segment with one lateral wax gland plate pair and a few posterior wax glands. IX abdominal segment bearing ovipositor. Costa, subcostal, and radius veins of forewing light green when alive. Second cubitus vein of forewing connected to radius vein while media and first cubitus veins disconnected or merely very weakly connected to radius vein. Cubitus vein of hindwing slightly curved and connected to radius vein while sometimes weakly disconnected (Fig. 7C).

Ecology. According to our collection, winged gallicolae mainly emerge from the gall during August to October, and we also recorded one gall sample of Dasyueshan whose winged gallicolae emerged in early December. We also found that A. breviacus, A. baborinisanensis, A. xueshanensis, and A. tsugae are sympatric in a certain region of Xueshan (121.280263, 24.38945).

Version of Record, first published online April 17, 2025, with fixed content and layout in compliance with Art. 8.1.3.2 ICZN.

Nagoya Protocol: The authors attest that all legal and regulatory requirements, including export and import collection permits, have been followed for the collection of specimens from source populations at any international, national, regional, or other geographic level for all relevant field specimens collected as part of this study.

Acknowledgements

The authors thank Daniel Lau for collecting adelgid specimens. This manuscript was edited by Wallace Academic Editing.

Author contributions

Ming-Hsun Chou (Data curation [lead], Formal analysis [lead], Investigation [lead], Methodology [lead], Visualization [lead], Writing—original draft [lead], Writing—review & editing [lead]), Zong-Yu Shen (Methodology [supporting], Visualization [supporting]), I-Hsuan Chu (Investigation [lead]), and Wen-Bin Yeh (Conceptualization [lead], Funding acquisition [lead], Investigation [lead], Resources [lead], Supervision [lead], Writing—review & editing [lead])

Funding

This study was supported by the Bureau of Animal and Plant Health Inspection and Quarantine, Council of Agriculture, Executive Yuan, Taiwan (grant numbers: 108AS-8.6.2-BQ-B1 and 109AS-8.1.2-BQ-B1).

Conflicts of interest. None declared.

References